There's a few issues here: - force-runtime does nothing, runtime is always vLLM - llmfit detecting 2GPUs with 48GB VRAM total, I only have a single 7900XTX [system.txt](https://github.com/user-attachments/files/26148277/system.txt) [rocminfo.txt](https://github.com/user-attachments/files/26148275/rocminfo.txt) [llmfit.txt](https://github.com/user-attachments/files/26148276/llmfit.txt)